Is there any data?

Yes. (high confidence | DPC, ISS)

Do numbers refer to ‘performed tests’ or ‘individuals tested’?

In Italy “tamponi” (“swabs”) refers to “performed tests”: many patients are tested several times (even 5 times). (high confidence | DPC)

In particular, patients who are recovered from COVID-19 are tested at least 3 times: 1 positive and 2 negative. (high confidence | DPC)

It should be taken into account the fact that “guariti” (“recovered”) in the official data refers not only to those actually recovered (who tested negative twice), but to all those discharged from hospital. (high confidence | DPC, IP, IS24O)

Are negative results included? Are pending results included?

The total number of performed tests includes negative test results. (high confidence | DPC)

The total number of performed tests includes those that are pending results. (low confidence | DPC, PM, RP, RL)

Do the figures include all tests conducted in the country, or only some?

Figures reported by Italy include all tests conducted in all working laboratories of the country. (medium confidence, MTP27-02, MTP09-03, MTP03-04)

Are all regions and laboratories within a country submitting data on the same basis?

Yes, all Italian regions and laboratories submit testing data on the same basis (i.e. “performed tests”). (medium-high confidence | DPC)

What period do the published figures refer to?

The figures published at a given date attempt to include all tests conducted up to that date, but significant and variable delays are present: both from symptom onset to lab confirmation (\(\Delta T_{symptoms}\)) and from lab confirmation to official database update (\(\Delta T_{lab}\)). (medium-high confidence | ISS)

For example, in Lombardy \(\Delta T_{symptoms}= 7.3 \text{ days } (1,20)_{95\% \text{CI}}\) and \(\Delta T_{lab} = 3.6 \text{ days } (1,10)_{95\% \text{CI}}\). (high confidence, AXV)

Are there any issues that affect the comparability of the data over time?

There have been significant changes over time in the ministerial testing protocol. (high confidence, ISS, MTP27-02, MTP09-03, MTP03-04, AXV)

What are the typical testing practices in the country?

Not all Italian regions follow the ministerial protocol of the central government, therefore there is significant regional heterogeneity in testing practices. (high confidence, IP, IP, IP)

Up to 20-03 we could identify 3 classes of testing practices described by the following pseudocode:

Consider two patients \(A\) and \(B\) living in the same house in the region \(R \in \{X, Y, Z\}\) such that

• \(A\) has no symptoms;
• \(B\) suffers from symptoms compatible with COVID-19;
• \(X\) respects the ministerial document dated 9-03 (e.g. Piedmont);
• \(Y\) makes more tests than those recommended by the ministerial document dated 9-03 (e.g. Veneto)
• \(Z\) makes fewer tests than those recommended by the ministerial document dated 9-03 (e.g. Lombardy).

If \(A\) and \(B\) live in \(X\):

• ASL sends an ambulance team of health-workers who perform the test on him, but not to \(A\) as he does not show any symptoms.
• \(A\) ends up in isolation, since in contact with a person who is definitely infected.

If \(A\) and \(B\) live in \(Y\):

• Test to \(B\)
• Test to \(A\)
• Test to the first and second neighbors (e.g. parents, friends, condominiums) in the hope of finding and isolating all the positives.

If \(A\) and \(B\) live in \(Z\):

• No test to \(B\) (unless \(B\) needs hospital care)
• No test to \(A\)
• No isolation from \(A\) because there is no evidence that \(B\) is positive despite its symptoms.

Might any of the information above be lost in translation?

Many important documents released by official sources are translated in English. (high confidence, DPC, ISS, MDS)